Obsolete
Status Update
No update yet.
Comments
dd...@chromium.org
What is the behavior of External Clear Key? Just modify the test to use that Key System name for Clear Key to rule out test differences.
jr...@chromium.org
Using External Clear Key with the drm-mp4-playback-temporary-multikey-sequential.html test fails due to "No supported PSSH box found", which is not surprising since the content doesn't have 'pssh' boxes for Common Encryption.
After tweaking the config (to pass 'keyids', just like the Clear Key version), and updating the message handler in drm-messagehandler.js (there is no license server to call, so just return the 'keyids' like the Clear Key version), the test runs the same as Clear Key.
Using the Clear Key version (clearkey-mp4-playback-temporary-multikey-sequential.html) with External Clear Key works the same as Clear Key.
Looking at the logs from chrome://media-internals, the delay talking to the license server may contribute to this. Will continue to investigate.
After tweaking the config (to pass 'keyids', just like the Clear Key version), and updating the message handler in drm-messagehandler.js (there is no license server to call, so just return the 'keyids' like the Clear Key version), the test runs the same as Clear Key.
Using the Clear Key version (clearkey-mp4-playback-temporary-multikey-sequential.html) with External Clear Key works the same as Clear Key.
Looking at the logs from chrome://media-internals, the delay talking to the license server may contribute to this. Will continue to investigate.
dd...@chromium.org
Interesting, that is not what I would have expected. (I thought the difference might be in the PPAPI path.) The delay is one additional difference. You could try delaying the Clear Key key "license" for a few seconds.
bu...@chromium.org
The following revision refers to this bug:
https://chromium.googlesource.com/chromium/src.git/+/21b0da050a97c275677ce0d15c143086fc3c0107
commit 21b0da050a97c275677ce0d15c143086fc3c0107
Author: jrummell <jrummell@chromium.org>
Date: Mon Sep 12 22:27:59 2016
media: Update logging to display the missing keyId
When playing encrypted media and playing stalls due to being unable
to decrypt the data, include the keyId missing in the logging output
(that includes chrome://media-internals).
BUG=645304
TEST=checked chrome://media-internals
Review-Url:https://codereview.chromium.org/2318933008
Cr-Commit-Position: refs/heads/master@{#418079}
[modify]https://crrev.com/21b0da050a97c275677ce0d15c143086fc3c0107/media/filters/decrypting_audio_decoder.cc
[modify]https://crrev.com/21b0da050a97c275677ce0d15c143086fc3c0107/media/filters/decrypting_demuxer_stream.cc
[modify]https://crrev.com/21b0da050a97c275677ce0d15c143086fc3c0107/media/filters/decrypting_video_decoder.cc
commit 21b0da050a97c275677ce0d15c143086fc3c0107
Author: jrummell <jrummell@chromium.org>
Date: Mon Sep 12 22:27:59 2016
media: Update logging to display the missing keyId
When playing encrypted media and playing stalls due to being unable
to decrypt the data, include the keyId missing in the logging output
(that includes chrome://media-internals).
BUG=645304
TEST=checked chrome://media-internals
Review-Url:
Cr-Commit-Position: refs/heads/master@{#418079}
[modify]
[modify]
[modify]
jr...@chromium.org
With additional logging, both External Clear Key and Widevine return the first 10 frames before returning status = kNoKey. However, ECK returns status = kSuccess with the first 10 frames, while WV returns status = kNeedMoreData. As a result, the frames returned by WV are never displayed (and thus the video never starts, and no 'stalled' event is raised later).
bu...@chromium.org
The following revision refers to this bug:
https://chromium.googlesource.com/chromium/src.git/+/e30a8fab234b1c9fa644f2053bd080620315536d
commit e30a8fab234b1c9fa644f2053bd080620315536d
Author: jrummell <jrummell@chromium.org>
Date: Sat Sep 17 02:43:32 2016
media: Update logging to indicate that new key has been provided
When playing encrypted media and decrypting stalls due to missing key,
add a message to indicate when a key is added and decryption tries to
start again.
BUG=645304
TEST=checked chrome://media-internals
Review-Url:https://codereview.chromium.org/2334953002
Cr-Commit-Position: refs/heads/master@{#419365}
[modify]https://crrev.com/e30a8fab234b1c9fa644f2053bd080620315536d/media/filters/decrypting_audio_decoder.cc
[modify]https://crrev.com/e30a8fab234b1c9fa644f2053bd080620315536d/media/filters/decrypting_demuxer_stream.cc
[modify]https://crrev.com/e30a8fab234b1c9fa644f2053bd080620315536d/media/filters/decrypting_video_decoder.cc
commit e30a8fab234b1c9fa644f2053bd080620315536d
Author: jrummell <jrummell@chromium.org>
Date: Sat Sep 17 02:43:32 2016
media: Update logging to indicate that new key has been provided
When playing encrypted media and decrypting stalls due to missing key,
add a message to indicate when a key is added and decryption tries to
start again.
BUG=645304
TEST=checked chrome://media-internals
Review-Url:
Cr-Commit-Position: refs/heads/master@{#419365}
[modify]
[modify]
[modify]
da...@chromium.org
This bug has been stale for more than a year. I've added the label StaleMediaBug and in two weeks will archive all such labeled bugs. If you want to keep it, please replace StaleMediaBug with StakeKeep. Thanks!
da...@chromium.org
Closing stale media bug as archived. Feel free to reopen if you think this was a mistake.
is...@google.com
This issue was migrated from crbug.com/chromium/645304?no_tracker_redirect=1
[Monorail components added to Component Tags custom field.]
[Monorail components added to Component Tags custom field.]
Description
ClearKey
- EME set up, provides key1.
- Video starts, displays a few frames, and freezes. Audio continues.
- After ~3 seconds, video element gets a 'stalled' event, which triggers the code to provide key2.
- Video resumes playing.
Widevine
- EME set up, provides key1.
- Video does not display anything.
WebKit allows the audio and video to get out of sync by up to 3 seconds [2]. This would explain the behavior for ClearKey.
Regardless whether ClearKey or Widevine is used, what happens for the test should be the same.
[1]
[2]